Tiling Assembly: a new tool for reference annotation-independent transcript assembly and novel gene identification by...

Area of Science:

• Genomics
• Bioinformatics
• Transcriptomics

Background:

• The annotation of the rice (Oryza sativa) genome remains incomplete despite advancements.
• Existing transcript assembly programs have limitations, including incomplete transcript identification and resource intensiveness.

Purpose of the Study:

• To evaluate and improve the effectiveness of transcript assembly programs for identifying novel genes in the rice genome.
• To develop a novel algorithm and pipeline to enhance genome annotation accuracy and efficiency.

Main Methods:

• RNA-sequencing (RNA-seq) data was analyzed using several transcript assembly programs, including Cufflinks.
• A new heuristic ab initio transcript assembly algorithm, Tiling Assembly, was developed.
• A pipeline was created to filter false-positive gene identifications, and Tiling Assembly was compared with Cufflinks.

Main Results:

• Cufflinks did not identify all transcripts present in the RNA-seq data.
• The Tiling Assembly algorithm demonstrated effectiveness in gene identification based on short read and junction alignment.
• Combining Tiling Assembly with Cufflinks identified 767 unannotated genes in the rice genome.
• Tiling Assembly accurately determined transcription start sites and proved to be organism-independent.

Conclusions:

• Combining Tiling Assembly and Cufflinks is highly efficient for novel gene identification in the rice genome.
• Tiling Assembly is a valuable, organism-independent tool for improving genome annotation accuracy.
• The developed pipeline effectively reduces false positives, enhancing the reliability of gene discovery.